Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
819528 | Composites Part B: Engineering | 2008 | 16 Pages |
Parametric variation and optimisation using genetic algorithms employing single and multi-objective functions are proposed for the optimisation of a structural steel/composite connection. The joint in marine applications is the connection between the steel hull and the composite superstructure of a naval vessel. A baseline joint is defined and all parametric variations and optimised joints are compared to this. The parametric results provided design curves of the joint performance determined from the weight, Von Mises stress in the adhesive and the global stiffness indicating performance sensitivity to specific changes in the joint geometry.The results indicated that the parametric variations can lead to an improvement in the performance but high levels of human interaction are required to make a combined improvement to the performance. The use of genetic algorithms provided an efficient method of searching the design space for an optimal joint. The single objective function provides an excellent reduction in the weight and maintaining or improving the performance of the joint to in-plane compressive loading. The use of the multi-objective function whereby a weighting was applied to the weight, stress and stiffness performance criteria proved extremely successful in further optimising the joint. The use of genetic algorithms has been demonstrated to efficiently search the complex design space of a structural connection and the use of multi-objective functions as the most effective selection method.